Disappearance of Gas in the Electric Discharge. 24 5> 



of simplicity is almost always misleading. For, in general,, 

 the current arriving at the electrodes is not at all simply 

 related to the processes of ionization and recombination 

 taking place in the gas, and it is on these processes that 

 a purely gaseous reaction is likely to depend, in particular, 

 it seems impossible that Faraday's law can ever be truly 

 obeyed tor a gaseous reaction proceeding in the independent 

 discharge without thermionic emission from the cathode ; 

 for in that discharge the emission of electrons from the 

 cathode is inseparably connected with the incidence of 

 positive ions upon it ; the current arriving at the cathode 

 must be made up of positive ions arriving and electrons 

 leaving, and these two parts of the current are likely to be 

 related in very different ways to the rate of any reaction 

 proceeding in the gas. 



15. A few experiments have also been made in H 2 and in 

 N 2 . In H 2 the disappearance of gas seems to be totally 

 independent of electrical conditions, as Langmuir, who first 

 described the phenomenon, suggests. It occurs when there 

 is no positive ionization, and its rate is unaltered even when 

 there is a strong field driving electrons (and negative ions) 

 back to the cathode. In N 2 (except when the temperature 

 of the tungsten filament reaches 2800° K.) the disappearance 

 of gas does not occur without ionization; again, — p is> 

 proportional to i p ; but now « appears independent of the 

 pressure. Such a difference is to be expected, for, if a 

 chemical reaction is involved in the disappearance, it is not 

 likely to be one between a charged and a neutral molecule- 

 Further, u is very much less than in CO ; the greatest value 

 observed was 0'3. The measurements were much less con- 

 sistent than in CO, the variation being probably due to the 

 fact that there nre several different processes causing the 

 disappearance of the gas_, the relative rates of which vary 

 according to the conditions. 



16. Last, a matter of a totally different nature may be 

 mentioned. In some experiments, especially those at pres- 

 sure less than O05 mm., the arguments discussed above 

 indicated that there was no recombination of the ions in the 

 gas ; ig/ie was independent within wide limits of i e , and a. was 

 almost exactly 2. And yet in those conditions the gas 

 emitted a bright glow consisting of the ordinary CO spec- 

 trum, so far as it can be identified by its characteristic lines 

 in the visual region. The modern theory of spectra suggests 

 that the complete spectrum is only developed when an 

 electron is removed to infinity and allowed to fall back : in 

 the conditions considered here, it is removed to infinity, but 



